Review of Tanker Safety after the Introduction of OPA 90

1. International Symposium of Maritime Safety Security & Environmental protection (MASSEP 2013) 30-31 May 2013 Athens, GREECE Review of Tanker Safety after the Introduction of OPA 90 Apostolos Papanikolaou Professor, Ship Design Laboratory, NTUA papa@deslab.ntua.gr, http://www.naval.ntua.gr/sdl Eleftheria Eliopoulou Dr Eng, Ship Design Laboratory, NTUA Rainer Hamann Dr Eng, Germanischer Lloyd SA

2. Contents • Introduction • Methodology of work • Analysis of results • Conclusion – The way ahead Review of Tanker Safety after the Introduction of OPA 90

3. Introduction The present study focuses on a comprehensive analysis of recorded accidents of medium and large crude oil tankers (DWT over 20,000 tonnes), which occurred after the introduction of OPA90 and until today. Raw casualty data was reviewed and re-analysed in order to produce appropriate statistics useful for the implementation of risk-based assessment methodologies. The main outcome of the presented study is the identification of significant historical trends and of quantitative characteristics of individual categories of tanker accidents, like overall accidental frequencies per ship year, frequencies of each major accident category and per tanker ship size, ship type/design and age, the degree of accidents’ severity and the oil spill tonne rates per ship year. Therefore this study is a valuable source of information for the assessment of the effectiveness of current IMO regulations, classification society rules and tanker industry’s practice. Review of Tanker Safety after the Introduction of OPA 90

4. Methodology of work (1) Field/Sampling Data Model In order to conduct a risk analysis assessment, historical casualty data was extracted from commercial casualty databases (IHS Fairplay, LMIU), imported in a new purposely designed database (NTUA-SDL), critically assessed and enhanced by other information publicly available (internet search, etc.). The particular step is considered of paramount importance for the reliability of the conducted risk analysis because: • Commercial databases were originally not designed for potential application in risk assessment procedures; • Their information is to a great extent available in textual form, whereas details of importance for formal risk assessment procedures (FSA) are missing. • In several cases, there was lack of or erratic information about principal issues for the analysis, namely on the consequences of the incident or/and on several steps of Event Tree analysis (missing or erroneous spillage extent for important and well publicised major tanker accidents). Review of Tanker Safety after the Introduction of OPA 90

5. Methodology of work (2) Raw data information Typical input data(frozen) from commercial databases (IHS-Fairplay, LMIU) Review of Tanker Safety after the Introduction of OPA 90

6. Methodology of work (3) NTUA-SDL database taba Typical main database entry form for the population of data in the NTUA-SDL database Review of Tanker Safety after the Introduction of OPA 90

7. Methodology of work (4) Sampling Plan The present study focuses on accidents that potentially lead to ship’s Loss Of Watertight Integrity (LOWI) and to accidental oil pollution, thus only the first six (6) categories of accidents are investigated as illustrated in the Figure (shaded box). Medium Crude Oil contains HANDYSIZE (20,000- 34,999 DWT) HANDYMAX (35,000 – 60,000 DWT) Large Crude Oil contains PANAMAX (60,000- 79,999 DWT) AFRAMAX (80,000 – 119,999 DWT) SUEZMAX (120,000 – 199,999 DWT) VLCC (200,000-319,999 DWT) ULCC (greater than 320,000 DWT) Review of Tanker Safety after the Introduction of OPA 90

8. Methodology of work (5) Operational Fleet at Risk Annual DH-ships population was very small in the first years of the analysis period, namely 1990-1995, but it steadily increased as could be expected because of the gradual enforcement of the requirements of double hull ship concept worldwide (and the corresponding phase out of single hull ships), surpassing for the first time the non-DH fleet in the period 2001-2003. Review of Tanker Safety after the Introduction of OPA 90

9. Analysis of Results (1) Frequencies of events Frequencies of serious and non-serious events per ship year with uncertainty margins (assumed binomial distribution, 95% confidence interval) Navigational accidents (collision, contact and grounding) as well as Non-accidental Structural Failures (NASF) exhibit significantly lower frequencies of serious cases, compared to the non-serious ones; this is confirmed for both the medium and large tankers. Fire and Explosion accidents appear to exhibit a different behaviour with respect to seriousness of accidents (with the recorded serious accidents prevailing), indicating that if such accident happen, higher probabilities of having serious consequences (and fatalities) should be expected. Review of Tanker Safety after the Introduction of OPA 90

10. Analysis of Results (2) Navigational accidents Focusing on the sampling plan of serious navigational accidents and assuming that accidental side damages are related to collisions and contacts, whereas bottom damages are attributed to grounding events, the following probabilities were calculated: Medium tankers: 57% side damages – 43% bottom damages, related frequencies: 4.62E-03 side, 3.51E-03 bottom. Large tankers: 51% side damages – 49% bottom damages, related frequencies: 4.06E-03 side, 3.90E-03 bottom. The above relationships should be compared to the 40% to 60% assumption of MARPOL’s reg. 23 (weight o f collisions vs. Grounding), which appears erratic and needs to be revisited. Review of Tanker Safety after the Introduction of OPA 90

11. Analysis of Results (3) Fire, & Explosion accidents Slightly decreasing tendency in the annual frequency of fire and explosion events along the studied period. Especially, in the second decade (after 1999), annual frequencies are confined within significantly smaller margins compared to the corresponding dates in the first decade of statistical analysis. It is believed that the introduction of the ISM Code, the stricter requirements for oil fuel lines’ protection (SOLAS Chapter II-2) that entered into force in 1998 for newbuildings, as well as the temperature sensing devices applied after 2002, have enhanced the safety of ships against fire and explosion events, as reflected in the above statistics. Review of Tanker Safety after the Introduction of OPA 90

12. Analysis of Results (4) NASF & the impact of ship’s age Large Double Hull Tankers, 22 non-accidental structural failures occurred over the studied period resulting to an average frequency of 1.8E-03; it is noted that 16 cases out of 22 registered NASF cases were reported for newly built ships, namely in the group age of 0-5 ship’s years old. Medium Double Hull Tankers present 12 occurred non-accidental structural failures yielding a frequency of 1.1E-03, which is actually lower than the corresponding one of large tankers and without any indication of higher frequency for the new-built ships. Review of Tanker Safety after the Introduction of OPA 90

13. Analysis of Results (5) NASF & the impact of ship’s age NASF and the impact of ship’s age Medium Tankers: it is found that there is a peak in the middle age (16-20 years) group, whereas frequencies for more aged DH ships (over 20 years) are lower (left hand diagram). The latter can be explained by the fact that there are only very few operating DH ships over 20 years of age, therefore the uncertainty of the particular result is rather high. Comparable results for large oil tankers show also complex (and may be unexpected) patterns of frequencies, namely high frequencies in the young age group (0-5 years), and then continuously increasing frequency after the 6-10 years of built. In the young age group (0-5 years) there is a remarkable structural failure rate for all large ship sub-sizes, except for VLCC/ULCC, (right hand diagram). Review of Tanker Safety after the Introduction of OPA 90

14. Analysis of Results (6) Ship’s Total Loss Medium tankers present higher frequencies for grounding, NASF and almost the same value in case of explosion events, compared to the corresponding ones derived from the large tankers analysis. Moreover, large tankers exhibit higher corresponding frequencies in cases of collision and fire accidents. These observations, however, need to be considered with caution, because of the large statistical uncertainty due to the scarcity of related historical data. Review of Tanker Safety after the Introduction of OPA 90

15. Analysis of Results (7) Marine Pollution For the investigated tanker ship sizes, it is trivially confirmed that the larger the ship the more severe is the environmental impact in case of accidental loss of watertight integrity. The figures present the oil released to the sea as a consequence of medium and large size oil tanker accidents during the studied time period. Review of Tanker Safety after the Introduction of OPA 90

16. Analysis of Results (8) Marine Pollution Spill Tonne Rate per ship year along with the 95% confidence intervals (C.I., mean obtained value). The spill tonne rates for large and medium tankers are not comparable, as could be expected, because of the large difference in the carried cargo capacity. It is acknowledged that the 95% confidence intervals with respect to the noted average values are quite wide, indicating that the spill tonne rates are decisively determined by a very limited number of individual accidents with large oil spills (catastrophic accidents), rather than by many small pollution events or average size. Proceeding to a more refined analysis for AFRAMAX tankers, the calculated oil released to the sea is 233,138 tonnes within the studied period (1990-2008). Taking into account that the AFRAMAX Fleet at Risk was calculated to be 9,786 shipyears and considering an average DWT value of 100,000 tonnes, also accounting for a ship’s life cycle of 25 years period, this leads to about 0.6% DWT rate of oil per AFRAMAX ship, which is expected to be released to the sea within the ship’s life cycle. This may be compared to the 0.015 (or 1.5% DWT) maximum allowable rate of oil spill for an AFRAMAX tanker according to MARPOL, reg23 [12], indicating that the MARPOL requirement is on the safe side. Review of Tanker Safety after the Introduction of OPA 90

17. Conclusions –The way ahead (1) The current study presented results of a systematic analysis of accidents pertaining to medium and large oil tankers (deadweight over 20 k tonnes) and covering the period after the introduction of OPA 90, namely 1990 to 2009 (October), continuing earlier studies of NTUA-SDL and Germanischer Lloyd on the design and safety of tankers. Calculated values derived from the statistics must be used with caution because available databases do not capture in general all accidents (problem of under-reporting), they partly include erratic information and provide always a snapshot of the status for a certain observation period. Thus, single accidents, when they happen, may have a significant impact on the accident frequencies and especially on the indentified consequences, especially when they are of ‘catastrophic’ character. In order to show the uncertainty of the frequencies presented in this paper a confidence analysis was also presented referring to the initiating main top events. The data in this paper provide the basis for the development of a risk model for medium size tankers, which complements earlier conducted studies for large tankers. Such a risk model should consider the uncertainty in the initial accident frequencies as well as in the dependent probabilities in the scenarios. This would allow considering the effect of uncertainty also in subsequent analyses, for instance in a cost-benefit analysis of design modifications. Review of Tanker Safety after the Introduction of OPA 90

18. Conclusions –The way ahead (2) Having completed earlier the Formal Safety Assessment procedure for Large Oil Tankers, the present study is the first one addressing systematically the effect of ship size in the frame of Formal Safety Assessment procedures providing information that will allow the development of more elaborated risk models taking into account the ship size. This is considered essential, particularly when dealing with ship types, for which their safety performance (here: safety of environment) may dramatically change with the increase of ship size. A comparison of the herein obtained results for the medium size tankers with corresponding ones for the large oil tankers enabled the identification of notable differences in the accidental pattern, though overall trends are comparable. Also, some important conclusions regarding relevant regulatory provisions of MARPOL regarding the accidental oil outflow index were enabled. In the next stage of this research, which will be complemented by the analysis of accidental data of the small tankers(deadweight below 20 k tons), the societal risk will be calculated and expressed by FN diagrams (denoting cumulative frequencies of losses of human lives) and FT diagrams (denoting cumulative frequencies of oil pollution in tonnes, i.e. the environmental impact by oil spillage); this will allow even more comprehensive conclusions on the safety of oil transport by all sizes of tankers. Review of Tanker Safety after the Introduction of OPA 90